5.7 hashtable
前面介绍的关联容器set、multiset、map和multimap的底层机制都是基于RB-Tree红黑树，虽然能够实现在插入、删除和搜索操作能够达到对数平均时间，可是要求输入数据有足够的随机性。本文介绍的hashtable不需要要求输入数据具有随机性，在插入、删除和搜索操作都能达到常数平均时间。本文介绍的hash
table是来自SGI STL中的<stl_hashtable.h>文件，在这里为了避免冲突（即不同元素映射到相同的键值位置），采用了拉链法来解决冲突问题。SGI中实现hash table的方式，在每个表格元素中维护一个链表, 然后在链表上执行元素的插入、搜寻、删除等操作，该表格中的每个元素被称为桶(bucket)。
哈希表的说明参考：http://blog.csdn.net/u012243115/article/details/40648687
hashtable散列表源码剖析
#ifndef __SGI_STL_INTERNAL_HASHTABLE_H
#define __SGI_STL_INTERNAL_HASHTABLE_H
// Hashtable class, used to implement the hashed associative containers
// hash_set, hash_map, hash_multiset, and hash_multimap.
//SGI STL的hashtable的实现方法是拉链法.
//拉链法可以避免hashtable的冲突问题,即不同数据映射到同一的hash值
#include <stl_algobase.h>
#include <stl_alloc.h>
#include <stl_construct.h>
#include <stl_tempbuf.h>
#include <stl_algo.h>
#include <stl_uninitialized.h>
#include <stl_function.h>
#include <stl_vector.h>
#include <stl_hash_fun.h>
__STL_BEGIN_NAMESPACE
//hashtable中链表的节点结构
//类似于单链表的节点结构
template <class _Val>
struct _Hashtable_node
{
_Hashtable_node* _M_next;//指向下一节点
_Val _M_val;//节点元素值
};
//这里使用前置声明, 避免在后面交叉引用会导致编译错误
template <class _Val, class _Key, class _HashFcn,
class _ExtractKey, class _EqualKey, class _Alloc = alloc>
class hashtable;
template <class _Val, class _Key, class _HashFcn,
class _ExtractKey, class _EqualKey, class _Alloc>
struct _Hashtable_iterator;
template <class _Val, class _Key, class _HashFcn,
class _ExtractKey, class _EqualKey, class _Alloc>
struct _Hashtable_const_iterator;
//hashtable迭代器定义
//注意：hash table迭代器没有提供后退操作operator--
//也没用提供逆向迭代器reverse iterator
template <class _Val, class _Key, class _HashFcn,
class _ExtractKey, class _EqualKey, class _Alloc>
struct _Hashtable_iterator {
//内嵌类型别名
typedef hashtable<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,_Alloc>
_Hashtable;
typedef _Hashtable_iterator<_Val, _Key, _HashFcn,
_ExtractKey, _EqualKey, _Alloc>
iterator;
typedef _Hashtable_const_iterator<_Val, _Key, _HashFcn,
_ExtractKey, _EqualKey, _Alloc>
const_iterator;
typedef _Hashtable_node<_Val> _Node;
typedef forward_iterator_tag iterator_category;//采用正向迭代器
typedef _Val value_type;
typedef ptrdiff_t difference_type;
typedef size_t size_type;
typedef _Val& reference;
typedef _Val* pointer;
_Node* _M_cur;//当前迭代器所指位置
_Hashtable* _M_ht;//hashtable中的位置,控制访问桶子连续性
_Hashtable_iterator(_Node* __n, _Hashtable* __tab)
: _M_cur(__n), _M_ht(__tab) {}
_Hashtable_iterator() {}
//返回当前节点元素值的引用
reference operator*() const { return _M_cur->_M_val; }
#ifndef __SGI_STL_NO_ARROW_OPERATOR
pointer operator->() const { return &(operator*()); }
#endif /* __SGI_STL_NO_ARROW_OPERATOR */
//操作符重载定义在后面定义
iterator& operator++();
iterator operator++(int);
//比较两个迭代器是否指向同一个节点
bool operator==(const iterator& __it) const
{ return _M_cur == __it._M_cur; }
bool operator!=(const iterator& __it) const
{ return _M_cur != __it._M_cur; }
};
//下面是const iterator的定义，基本和上面相同
template <class _Val, class _Key, class _HashFcn,
class _ExtractKey, class _EqualKey, class _Alloc>
struct _Hashtable_const_iterator {
typedef hashtable<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,_Alloc>
_Hashtable;
typedef _Hashtable_iterator<_Val,_Key,_HashFcn,
_ExtractKey,_EqualKey,_Alloc>
iterator;
typedef _Hashtable_const_iterator<_Val, _Key, _HashFcn,
_ExtractKey, _EqualKey, _Alloc>
const_iterator;
typedef _Hashtable_node<_Val> _Node;
typedef forward_iterator_tag iterator_category;
typedef _Val value_type;
typedef ptrdiff_t difference_type;
typedef size_t size_type;
typedef const _Val& reference;
typedef const _Val* pointer;
const _Node* _M_cur;
const _Hashtable* _M_ht;
_Hashtable_const_iterator(const _Node* __n, const _Hashtable* __tab)
: _M_cur(__n), _M_ht(__tab) {}
_Hashtable_const_iterator() {}
_Hashtable_const_iterator(const iterator& __it)
: _M_cur(__it._M_cur), _M_ht(__it._M_ht) {}
reference operator*() const { return _M_cur->_M_val; }
#ifndef __SGI_STL_NO_ARROW_OPERATOR
pointer operator->() const { return &(operator*()); }
#endif /* __SGI_STL_NO_ARROW_OPERATOR */
const_iterator& operator++();
const_iterator operator++(int);
bool operator==(const const_iterator& __it) const
{ return _M_cur == __it._M_cur; }
bool operator!=(const const_iterator& __it) const
{ return _M_cur != __it._M_cur; }
};
// Note: assumes long is at least 32 bits.
// 注意：假设long至少为32-bits, 可以根据自己需要修改
//定义28个素数用作hashtable的大小
enum { __stl_num_primes = 28 };
static const unsigned long __stl_prime_list[__stl_num_primes] =
{
53ul, 97ul, 193ul, 389ul, 769ul,
1543ul, 3079ul, 6151ul, 12289ul, 24593ul,
49157ul, 98317ul, 196613ul, 393241ul, 786433ul,
1572869ul, 3145739ul, 6291469ul, 12582917ul, 25165843ul,
50331653ul, 100663319ul, 201326611ul, 402653189ul, 805306457ul,
1610612741ul, 3221225473ul, 4294967291ul
};
//返回大于n的最小素数
inline unsigned long __stl_next_prime(unsigned long __n)
{
const unsigned long* __first = __stl_prime_list;
const unsigned long* __last = __stl_prime_list + (int)__stl_num_primes;
const unsigned long* pos = lower_bound(__first, __last, __n);
/*
上面的lower_bound调用的是STL中的算法lower_bound();
该算法的功能：
Returns an iterator pointing to the first element in the range [first,last) which does not compare less than val.
The elements in the range shall already be sorted according to this same criterion (operator< or comp)
即返回在[first,last)范围内第一个不小于val的位置
注意：调用该算法之前，[first,last)范围里面的元素必须已排序
该算法的原型如下：
第一个版本：采用默认比较准则operator<
template <class ForwardIterator, class T>
ForwardIterator lower_bound (ForwardIterator first, ForwardIterator last,
const T& val);
第二版本：采用用户指定的比较函数comp
template <class ForwardIterator, class T, class Compare>
ForwardIterator lower_bound (ForwardIterator first, ForwardIterator last,
const T& val, Compare comp);
其实该算法的实现机制使用二分查找法进行查找元素val：
template <class ForwardIterator, class T>
ForwardIterator lower_bound (ForwardIterator first, ForwardIterator last, const T& val)
{
ForwardIterator it;
iterator_traits<ForwardIterator>::difference_type count, step;
count = distance(first,last);
while (count>0)
{
it = first; step=count/2; advance (it,step);
if (*it<val) { // or: if (comp(*it,val)), for version (2)
first=++it;
count-=step+1;
}
else count=step;
}
return first;
}
下面给出例子：lower_bound/upper_bound example
#include <iostream> // std::cout
#include <algorithm> // std::lower_bound, std::upper_bound, std::sort
#include <vector> // std::vector
int main () {
int myints[] = {10,20,30,30,20,10,10,20};
std::vector<int> v(myints,myints+8); // 10 20 30 30 20 10 10 20
std::sort (v.begin(), v.end()); // 10 10 10 20 20 20 30 30
std::vector<int>::iterator low,up;
low=std::lower_bound (v.begin(), v.end(), 20); // ^
up= std::upper_bound (v.begin(), v.end(), 20); // ^
std::cout << "lower_bound at position " << (low- v.begin()) << '\n';
std::cout << "upper_bound at position " << (up - v.begin()) << '\n';
return 0;
}
Output:
lower_bound at position 3
upper_bound at position 6
*/
return pos == __last ? *(__last - 1) : *pos;
}
// Forward declaration of operator==.
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
class hashtable;
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
bool operator==(const hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>& __ht1,
const hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>& __ht2);
// Hashtables handle allocators a bit differently than other containers
// do. If we're using standard-conforming allocators, then a hashtable
// unconditionally has a member variable to hold its allocator, even if
// it so happens that all instances of the allocator type are identical.
// This is because, for hashtables, this extra storage is negligible.
// Additionally, a base class wouldn't serve any other purposes; it
// wouldn't, for example, simplify the exception-handling code.
//hash table的定义
//模板参数定义
/*
Value： 节点的实值类型
Key： 节点的键值类型
HashFcn： hash function的类型
ExtractKey：从节点中取出键值的方法（函数或仿函数）
EqualKey：判断键值是否相同的方法（函数或仿函数）
Alloc：空间配置器
*/
//hash table的线性表是用vector容器维护
template <class _Val, class _Key, class _HashFcn,
class _ExtractKey, class _EqualKey, class _Alloc>
class hashtable {
public:
typedef _Key key_type;
typedef _Val value_type;
typedef _HashFcn hasher;
typedef _EqualKey key_equal;
typedef size_t size_type;
typedef ptrdiff_t difference_type;
typedef value_type* pointer;
typedef const value_type* const_pointer;
typedef value_type& reference;
typedef const value_type& const_reference;
hasher hash_funct() const { return _M_hash; }
key_equal key_eq() const { return _M_equals; }
private:
typedef _Hashtable_node<_Val> _Node;
#ifdef __STL_USE_STD_ALLOCATORS
public:
typedef typename _Alloc_traits<_Val,_Alloc>::allocator_type allocator_type;
allocator_type get_allocator() const { return _M_node_allocator; }
private:
typename _Alloc_traits<_Node, _Alloc>::allocator_type _M_node_allocator;
_Node* _M_get_node() { return _M_node_allocator.allocate(1); }
void _M_put_node(_Node* __p) { _M_node_allocator.deallocate(__p, 1); }
# define __HASH_ALLOC_INIT(__a) _M_node_allocator(__a),
#else /* __STL_USE_STD_ALLOCATORS */
public:
typedef _Alloc allocator_type;
allocator_type get_allocator() const { return allocator_type(); }
private:
typedef simple_alloc<_Node, _Alloc> _M_node_allocator_type;
_Node* _M_get_node() { return _M_node_allocator_type::allocate(1); }
void _M_put_node(_Node* __p) { _M_node_allocator_type::deallocate(__p, 1); }
# define __HASH_ALLOC_INIT(__a)
#endif /* __STL_USE_STD_ALLOCATORS */
//以下是hash table的成员变量
private:
hasher _M_hash;
key_equal _M_equals;
_ExtractKey _M_get_key;
vector<_Node*,_Alloc> _M_buckets;//用vector维护buckets
size_type _M_num_elements;//hashtable中list节点个数
public:
typedef _Hashtable_iterator<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,_Alloc>
iterator;
typedef _Hashtable_const_iterator<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,
_Alloc>
const_iterator;
friend struct
_Hashtable_iterator<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,_Alloc>;
friend struct
_Hashtable_const_iterator<_Val,_Key,_HashFcn,_ExtractKey,_EqualKey,_Alloc>;
public:
//构造函数
hashtable(size_type __n,
const _HashFcn& __hf,
const _EqualKey& __eql,
const _ExtractKey& __ext,
const allocator_type& __a = allocator_type())
: __HASH_ALLOC_INIT(__a)
_M_hash(__hf),
_M_equals(__eql),
_M_get_key(__ext),
_M_buckets(__a),
_M_num_elements(0)
{
_M_initialize_buckets(__n);//预留空间,并将其初始化为空0
//预留空间大小为大于n的最小素数
}
hashtable(size_type __n,
const _HashFcn& __hf,
const _EqualKey& __eql,
const allocator_type& __a = allocator_type())
: __HASH_ALLOC_INIT(__a)
_M_hash(__hf),
_M_equals(__eql),
_M_get_key(_ExtractKey()),
_M_buckets(__a),
_M_num_elements(0)
{
_M_initialize_buckets(__n);
}
//拷贝构造函数
hashtable(const hashtable& __ht)
: __HASH_ALLOC_INIT(__ht.get_allocator())
_M_hash(__ht._M_hash),
_M_equals(__ht._M_equals),
_M_get_key(__ht._M_get_key),
_M_buckets(__ht.get_allocator()),
_M_num_elements(0)
{
_M_copy_from(__ht);//复制hashtable内容
}
#undef __HASH_ALLOC_INIT
//可以通过operator=初始化hashtable对象
hashtable& operator= (const hashtable& __ht)
{
if (&__ht != this) {
clear();
_M_hash = __ht._M_hash;
_M_equals = __ht._M_equals;
_M_get_key = __ht._M_get_key;
_M_copy_from(__ht);
}
return *this;
}
~hashtable() { clear(); }
//返回hashtable元素的个数
size_type size() const { return _M_num_elements; }
size_type max_size() const { return size_type(-1); }
bool empty() const { return size() == 0; }
//交换两个hashtable的内容
void swap(hashtable& __ht)
{
__STD::swap(_M_hash, __ht._M_hash);
__STD::swap(_M_equals, __ht._M_equals);
__STD::swap(_M_get_key, __ht._M_get_key);
_M_buckets.swap(__ht._M_buckets);
__STD::swap(_M_num_elements, __ht._M_num_elements);
}
iterator begin()
{
for (size_type __n = 0; __n < _M_buckets.size(); ++__n)
if (_M_buckets[__n])//若hashtable中的桶子_M_buckets有链表list
return iterator(_M_buckets[__n], this);//返回链表的第一个节点位置
return end();//若list链表为空,则返回尾端end(),其实这里尾端和起始端一样
}
iterator end() { return iterator(0, this); }//返回桶子list链表的null指针,即尾端
const_iterator begin() const
{
for (size_type __n = 0; __n < _M_buckets.size(); ++__n)
if (_M_buckets[__n])
return const_iterator(_M_buckets[__n], this);
return end();
}
const_iterator end() const { return const_iterator(0, this); }
#ifdef __STL_MEMBER_TEMPLATES
template <class _Vl, class _Ky, class _HF, class _Ex, class _Eq, class _Al>
friend bool operator== (const hashtable<_Vl, _Ky, _HF, _Ex, _Eq, _Al>&,
const hashtable<_Vl, _Ky, _HF, _Ex, _Eq, _Al>&);
#else /* __STL_MEMBER_TEMPLATES */
friend bool __STD_QUALIFIER
operator== __STL_NULL_TMPL_ARGS (const hashtable&, const hashtable&);
#endif /* __STL_MEMBER_TEMPLATES */
public:
//返回桶子个数,即线性表中节点数
size_type bucket_count() const { return _M_buckets.size(); }
//线性表最多分配节点数
size_type max_bucket_count() const
{ return __stl_prime_list[(int)__stl_num_primes - 1]; }
//指定桶子键值key中list链表的元素个数
size_type elems_in_bucket(size_type __bucket) const
{
size_type __result = 0;
for (_Node* __cur = _M_buckets[__bucket]; __cur; __cur = __cur->_M_next)
__result += 1;
return __result;
}
//插入元素节点,不允许存在重复元素
pair<iterator, bool> insert_unique(const value_type& __obj)
{
//判断容量是否够用, 否则就重新配置
resize(_M_num_elements + 1);
//插入元素,不允许存在重复元素
return insert_unique_noresize(__obj);
}
//插入元素节点,允许存在重复元素
iterator insert_equal(const value_type& __obj)
{//判断容量是否够用, 否则就重新配置
resize(_M_num_elements + 1);
//插入元素,允许存在重复元素
return insert_equal_noresize(__obj);
}
//具体定义见hashtable类外后面的剖析
pair<iterator, bool> insert_unique_noresize(const value_type& __obj);
iterator insert_equal_noresize(const value_type& __obj);
#ifdef __STL_MEMBER_TEMPLATES
template <class _InputIterator>
void insert_unique(_InputIterator __f, _InputIterator __l)
{
insert_unique(__f, __l, __ITERATOR_CATEGORY(__f));
}
template <class _InputIterator>
void insert_equal(_InputIterator __f, _InputIterator __l)
{
insert_equal(__f, __l, __ITERATOR_CATEGORY(__f));
}
template <class _InputIterator>
void insert_unique(_InputIterator __f, _InputIterator __l,
input_iterator_tag)
{
for ( ; __f != __l; ++__f)
insert_unique(*__f);
}
template <class _InputIterator>
void insert_equal(_InputIterator __f, _InputIterator __l,
input_iterator_tag)
{
for ( ; __f != __l; ++__f)
insert_equal(*__f);
}
template <class _ForwardIterator>
void insert_unique(_ForwardIterator __f, _ForwardIterator __l,
forward_iterator_tag)
{
size_type __n = 0;
distance(__f, __l, __n);
resize(_M_num_elements + __n);
for ( ; __n > 0; --__n, ++__f)
insert_unique_noresize(*__f);
}
template <class _ForwardIterator>
void insert_equal(_ForwardIterator __f, _ForwardIterator __l,
forward_iterator_tag)
{
size_type __n = 0;
distance(__f, __l, __n);
resize(_M_num_elements + __n);
for ( ; __n > 0; --__n, ++__f)
insert_equal_noresize(*__f);
}
#else /* __STL_MEMBER_TEMPLATES */
void insert_unique(const value_type* __f, const value_type* __l)
{
size_type __n = __l - __f;
resize(_M_num_elements + __n);
for ( ; __n > 0; --__n, ++__f)
insert_unique_noresize(*__f);
}
void insert_equal(const value_type* __f, const value_type* __l)
{
size_type __n = __l - __f;
resize(_M_num_elements + __n);
for ( ; __n > 0; --__n, ++__f)
insert_equal_noresize(*__f);
}
void insert_unique(const_iterator __f, const_iterator __l)
{
size_type __n = 0;
distance(__f, __l, __n);
resize(_M_num_elements + __n);
for ( ; __n > 0; --__n, ++__f)
insert_unique_noresize(*__f);
}
void insert_equal(const_iterator __f, const_iterator __l)
{
size_type __n = 0;
distance(__f, __l, __n);
resize(_M_num_elements + __n);
for ( ; __n > 0; --__n, ++__f)
insert_equal_noresize(*__f);
}
#endif /*__STL_MEMBER_TEMPLATES */
reference find_or_insert(const value_type& __obj);
//查找指定键值的元素
iterator find(const key_type& __key)
{
size_type __n = _M_bkt_num_key(__key);//获取键值
_Node* __first;
for ( __first = _M_buckets[__n];
__first && !_M_equals(_M_get_key(__first->_M_val), __key);
__first = __first->_M_next)
{}
return iterator(__first, this);
}
const_iterator find(const key_type& __key) const
{
size_type __n = _M_bkt_num_key(__key);
const _Node* __first;
for ( __first = _M_buckets[__n];
__first && !_M_equals(_M_get_key(__first->_M_val), __key);
__first = __first->_M_next)
{}
return const_iterator(__first, this);
}
//返回键值为key的元素的个数
size_type count(const key_type& __key) const
{
const size_type __n = _M_bkt_num_key(__key);
size_type __result = 0;
for (const _Node* __cur = _M_buckets[__n]; __cur; __cur = __cur->_M_next)
if (_M_equals(_M_get_key(__cur->_M_val), __key))
++__result;
return __result;
}
pair<iterator, iterator>
equal_range(const key_type& __key);
pair<const_iterator, const_iterator>
equal_range(const key_type& __key) const;
//擦除元素
size_type erase(const key_type& __key);
void erase(const iterator& __it);
void erase(iterator __first, iterator __last);
void erase(const const_iterator& __it);
void erase(const_iterator __first, const_iterator __last);
void resize(size_type __num_elements_hint);
void clear();
private:
//返回大于n的最小素数
//实际上调用__stl_next_prime(__n);
size_type _M_next_size(size_type __n) const
{ return __stl_next_prime(__n); }
//预留空间,并将其初始化为0
void _M_initialize_buckets(size_type __n)
{
//返回大于n的最小素数__n_buckets
const size_type __n_buckets = _M_next_size(__n);
//这里调用vector的成员函数reserve
//reserve该函数功能是改变可用空间的大小
//Requests that the vector capacity be at least enough to contain __n_buckets elements.
_M_buckets.reserve(__n_buckets);
//调用vector的插入函数insert
//在原始end后面连续插入__n_buckets个0
_M_buckets.insert(_M_buckets.end(), __n_buckets, (_Node*) 0);
_M_num_elements = 0;
}
//获取键值key在桶子的位置
size_type _M_bkt_num_key(const key_type& __key) const
{
return _M_bkt_num_key(__key, _M_buckets.size());
}
//获取在桶子的序号，也就是键值
//输入参数是实值value
size_type _M_bkt_num(const value_type& __obj) const
{
return _M_bkt_num_key(_M_get_key(__obj));
}
size_type _M_bkt_num_key(const key_type& __key, size_t __n) const
{
return _M_hash(__key) % __n;//采用除法取余hash函数
}
size_type _M_bkt_num(const value_type& __obj, size_t __n) const
{
return _M_bkt_num_key(_M_get_key(__obj), __n);
}
//分配节点空间，并构造对象
_Node* _M_new_node(const value_type& __obj)
{
_Node* __n = _M_get_node();
__n->_M_next = 0;
__STL_TRY {
construct(&__n->_M_val, __obj);
return __n;
}
__STL_UNWIND(_M_put_node(__n));
}
//析构对象，并释放空间
void _M_delete_node(_Node* __n)
{
destroy(&__n->_M_val);
_M_put_node(__n);
}
void _M_erase_bucket(const size_type __n, _Node* __first, _Node* __last);
void _M_erase_bucket(const size_type __n, _Node* __last);
void _M_copy_from(const hashtable& __ht);
};
//前缀operator++重载，前进一个list节点
template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
class _All>
_Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>&
_Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>::operator++()
{
const _Node* __old = _M_cur;
_M_cur = _M_cur->_M_next;//若存在，则返回
//若当前节点为空，则需前进到下一个桶子的节点
if (!_M_cur) {
//根据元素值，定位出下一个bucket的位置，其起始位置就是我们的目的地
size_type __bucket = _M_ht->_M_bkt_num(__old->_M_val);
while (!_M_cur && ++__bucket < _M_ht->_M_buckets.size())
_M_cur = _M_ht->_M_buckets[__bucket];
}
return *this;
}
//后缀operator++重载
template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
class _All>
inline _Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>
_Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>::operator++(int)
{
iterator __tmp = *this;
++*this;//调用operator++
return __tmp;
}
template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
class _All>
_Hashtable_const_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>&
_Hashtable_const_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>::operator++()
{
const _Node* __old = _M_cur;
_M_cur = _M_cur->_M_next;
if (!_M_cur) {
size_type __bucket = _M_ht->_M_bkt_num(__old->_M_val);
while (!_M_cur && ++__bucket < _M_ht->_M_buckets.size())
_M_cur = _M_ht->_M_buckets[__bucket];
}
return *this;
}
template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
class _All>
inline _Hashtable_const_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>
_Hashtable_const_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>::operator++(int)
{
const_iterator __tmp = *this;
++*this;
return __tmp;
}
#ifndef __STL_CLASS_PARTIAL_SPECIALIZATION
template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
class _All>
inline forward_iterator_tag
iterator_category(const _Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>&)
{
return forward_iterator_tag();
}
template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
class _All>
inline _Val*
value_type(const _Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>&)
{
return (_Val*) 0;
}
template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
class _All>
inline hashtable<_Val,_Key,_HF,_ExK,_EqK,_All>::difference_type*
distance_type(const _Hashtable_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>&)
{
return (hashtable<_Val,_Key,_HF,_ExK,_EqK,_All>::difference_type*) 0;
}
template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
class _All>
inline forward_iterator_tag
iterator_category(const _Hashtable_const_iterator<_Val,_Key,_HF,
_ExK,_EqK,_All>&)
{
return forward_iterator_tag();
}
template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
class _All>
inline _Val*
value_type(const _Hashtable_const_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>&)
{
return (_Val*) 0;
}
template <class _Val, class _Key, class _HF, class _ExK, class _EqK,
class _All>
inline hashtable<_Val,_Key,_HF,_ExK,_EqK,_All>::difference_type*
distance_type(const _Hashtable_const_iterator<_Val,_Key,_HF,_ExK,_EqK,_All>&)
{
return (hashtable<_Val,_Key,_HF,_ExK,_EqK,_All>::difference_type*) 0;
}
#endif /* __STL_CLASS_PARTIAL_SPECIALIZATION */
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
bool operator==(const hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>& __ht1,
const hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>& __ht2)
{
typedef typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::_Node _Node;
if (__ht1._M_buckets.size() != __ht2._M_buckets.size())
return false;
for (int __n = 0; __n < __ht1._M_buckets.size(); ++__n) {
_Node* __cur1 = __ht1._M_buckets[__n];
_Node* __cur2 = __ht2._M_buckets[__n];
for ( ; __cur1 && __cur2 && __cur1->_M_val == __cur2->_M_val;
__cur1 = __cur1->_M_next, __cur2 = __cur2->_M_next)
{}
if (__cur1 || __cur2)
return false;
}
return true;
}
#ifdef __STL_FUNCTION_TMPL_PARTIAL_ORDER
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
inline bool operator!=(const hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>& __ht1,
const hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>& __ht2) {
return !(__ht1 == __ht2);
}
template <class _Val, class _Key, class _HF, class _Extract, class _EqKey,
class _All>
inline void swap(hashtable<_Val, _Key, _HF, _Extract, _EqKey, _All>& __ht1,
hashtable<_Val, _Key, _HF, _Extract, _EqKey, _All>& __ht2) {
__ht1.swap(__ht2);
}
#endif /* __STL_FUNCTION_TMPL_PARTIAL_ORDER */
//插入元素，不需要重新调整内存空间,不允许存在重复元素
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
pair<typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::iterator, bool>
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
::insert_unique_noresize(const value_type& __obj)
{
//获取待插入元素在hashtable中的桶子位置
const size_type __n = _M_bkt_num(__obj);
_Node* __first = _M_buckets[__n];
//判断hashtable中是否存在与之相等的键值元素
//若存在则不插入
//否则插入该元素
for (_Node* __cur = __first; __cur; __cur = __cur->_M_next)
if (_M_equals(_M_get_key(__cur->_M_val), _M_get_key(__obj)))
return pair<iterator, bool>(iterator(__cur, this), false);
//把元素插入到第一个节点位置
_Node* __tmp = _M_new_node(__obj);
__tmp->_M_next = __first;
_M_buckets[__n] = __tmp;
++_M_num_elements;
return pair<iterator, bool>(iterator(__tmp, this), true);
}
//插入元素，允许重复,不需要分配新的空间
//也就是说有足够的空间
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::iterator
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
::insert_equal_noresize(const value_type& __obj)
{
//获取待插入元素在hashtable中的桶子位置
const size_type __n = _M_bkt_num(__obj);
_Node* __first = _M_buckets[__n];
for (_Node* __cur = __first; __cur; __cur = __cur->_M_next)
//若存在键值相同的元素，则插在相同元素下一个位置
if (_M_equals(_M_get_key(__cur->_M_val), _M_get_key(__obj))) {
_Node* __tmp = _M_new_node(__obj);//创建新节点
__tmp->_M_next = __cur->_M_next;//将新节点插在当前节点之后
__cur->_M_next = __tmp;
++_M_num_elements;//节点数加1
return iterator(__tmp, this);//返回指向新增节点迭代器
}
//若不存在相同键值的元素,则插在第一个位置
_Node* __tmp = _M_new_node(__obj);//创建新节点
__tmp->_M_next = __first;//插入在链表表头
_M_buckets[__n] = __tmp;
++_M_num_elements;//节点数加1
return iterator(__tmp, this);//返回指向新增节点的迭代器
}
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::reference
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::find_or_insert(const value_type& __obj)
{
resize(_M_num_elements + 1);
size_type __n = _M_bkt_num(__obj);
_Node* __first = _M_buckets[__n];
for (_Node* __cur = __first; __cur; __cur = __cur->_M_next)
if (_M_equals(_M_get_key(__cur->_M_val), _M_get_key(__obj)))
return __cur->_M_val;
_Node* __tmp = _M_new_node(__obj);
__tmp->_M_next = __first;
_M_buckets[__n] = __tmp;
++_M_num_elements;
return __tmp->_M_val;
}
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
pair<typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::iterator,
typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::iterator>
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::equal_range(const key_type& __key)
{
typedef pair<iterator, iterator> _Pii;
const size_type __n = _M_bkt_num_key(__key);
for (_Node* __first = _M_buckets[__n]; __first; __first = __first->_M_next)
if (_M_equals(_M_get_key(__first->_M_val), __key)) {
for (_Node* __cur = __first->_M_next; __cur; __cur = __cur->_M_next)
if (!_M_equals(_M_get_key(__cur->_M_val), __key))
return _Pii(iterator(__first, this), iterator(__cur, this));
for (size_type __m = __n + 1; __m < _M_buckets.size(); ++__m)
if (_M_buckets[__m])
return _Pii(iterator(__first, this),
iterator(_M_buckets[__m], this));
return _Pii(iterator(__first, this), end());
}
return _Pii(end(), end());
}
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
pair<typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::const_iterator,
typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::const_iterator>
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
::equal_range(const key_type& __key) const
{
typedef pair<const_iterator, const_iterator> _Pii;
const size_type __n = _M_bkt_num_key(__key);
for (const _Node* __first = _M_buckets[__n] ;
__first;
__first = __first->_M_next) {
if (_M_equals(_M_get_key(__first->_M_val), __key)) {
for (const _Node* __cur = __first->_M_next;
__cur;
__cur = __cur->_M_next)
if (!_M_equals(_M_get_key(__cur->_M_val), __key))
return _Pii(const_iterator(__first, this),
const_iterator(__cur, this));
for (size_type __m = __n + 1; __m < _M_buckets.size(); ++__m)
if (_M_buckets[__m])
return _Pii(const_iterator(__first, this),
const_iterator(_M_buckets[__m], this));
return _Pii(const_iterator(__first, this), end());
}
}
return _Pii(end(), end());
}
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
typename hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::size_type
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::erase(const key_type& __key)
{
const size_type __n = _M_bkt_num_key(__key);
_Node* __first = _M_buckets[__n];
size_type __erased = 0;
if (__first) {
_Node* __cur = __first;
_Node* __next = __cur->_M_next;
while (__next) {
if (_M_equals(_M_get_key(__next->_M_val), __key)) {
__cur->_M_next = __next->_M_next;
_M_delete_node(__next);
__next = __cur->_M_next;
++__erased;
--_M_num_elements;
}
else {
__cur = __next;
__next = __cur->_M_next;
}
}
if (_M_equals(_M_get_key(__first->_M_val), __key)) {
_M_buckets[__n] = __first->_M_next;
_M_delete_node(__first);
++__erased;
--_M_num_elements;
}
}
return __erased;
}
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::erase(const iterator& __it)
{
_Node* __p = __it._M_cur;
if (__p) {
const size_type __n = _M_bkt_num(__p->_M_val);
_Node* __cur = _M_buckets[__n];
if (__cur == __p) {
_M_buckets[__n] = __cur->_M_next;
_M_delete_node(__cur);
--_M_num_elements;
}
else {
_Node* __next = __cur->_M_next;
while (__next) {
if (__next == __p) {
__cur->_M_next = __next->_M_next;
_M_delete_node(__next);
--_M_num_elements;
break;
}
else {
__cur = __next;
__next = __cur->_M_next;
}
}
}
}
}
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
::erase(iterator __first, iterator __last)
{
size_type __f_bucket = __first._M_cur ?
_M_bkt_num(__first._M_cur->_M_val) : _M_buckets.size();
size_type __l_bucket = __last._M_cur ?
_M_bkt_num(__last._M_cur->_M_val) : _M_buckets.size();
if (__first._M_cur == __last._M_cur)
return;
else if (__f_bucket == __l_bucket)
_M_erase_bucket(__f_bucket, __first._M_cur, __last._M_cur);
else {
_M_erase_bucket(__f_bucket, __first._M_cur, 0);
for (size_type __n = __f_bucket + 1; __n < __l_bucket; ++__n)
_M_erase_bucket(__n, 0);
if (__l_bucket != _M_buckets.size())
_M_erase_bucket(__l_bucket, __last._M_cur);
}
}
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
inline void
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::erase(const_iterator __first,
const_iterator __last)
{
erase(iterator(const_cast<_Node*>(__first._M_cur),
const_cast<hashtable*>(__first._M_ht)),
iterator(const_cast<_Node*>(__last._M_cur),
const_cast<hashtable*>(__last._M_ht)));
}
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
inline void
hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::erase(const const_iterator& __it)
{
erase(iterator(const_cast<_Node*>(__it._M_cur),
const_cast<hashtable*>(__it._M_ht)));
}
//调整hashtable的容量
//新的容量大小为__num_elements_hint
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
::resize(size_type __num_elements_hint)
{
//hashtable原始大小
const size_type __old_n = _M_buckets.size();
if (__num_elements_hint > __old_n) {//若新的容量大小比原始的大
//查找不低于__num_elements_hint的最小素数
const size_type __n = _M_next_size(__num_elements_hint);
if (__n > __old_n) {
//创建新的线性表,容量为__n，只是起到中介作用
vector<_Node*, _All> __tmp(__n, (_Node*)(0),
_M_buckets.get_allocator());
__STL_TRY {//以下是复制数据
for (size_type __bucket = 0; __bucket < __old_n; ++__bucket) {
_Node* __first = _M_buckets[__bucket];
while (__first) {
//获取实值在新桶子的键值位置
size_type __new_bucket = _M_bkt_num(__first->_M_val, __n);
//这个只是为了方便while循环里面__first的迭代
_M_buckets[__bucket] = __first->_M_next;
//将当前节点插入到新的桶子__new_bucket里面,成为list的第一个节点
__first->_M_next = __tmp[__new_bucket];//__first->_M_next指向null指针,因为新桶子是空的
__tmp[__new_bucket] = __first;//新桶子对应键值指向第一个节点
__first = _M_buckets[__bucket];//更新当前指针
}
}
_M_buckets.swap(__tmp);//交换内容
}
# ifdef __STL_USE_EXCEPTIONS
catch(...) {//释放临时hashtable的线性表tmp
for (size_type __bucket = 0; __bucket < __tmp.size(); ++__bucket) {
while (__tmp[__bucket]) {
_Node* __next = __tmp[__bucket]->_M_next;
_M_delete_node(__tmp[__bucket]);
__tmp[__bucket] = __next;
}
}
throw;
}
# endif /* __STL_USE_EXCEPTIONS */
}
}
}
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
::_M_erase_bucket(const size_type __n, _Node* __first, _Node* __last)
{
_Node* __cur = _M_buckets[__n];
if (__cur == __first)
_M_erase_bucket(__n, __last);
else {
_Node* __next;
for (__next = __cur->_M_next;
__next != __first;
__cur = __next, __next = __cur->_M_next)
;
while (__next != __last) {
__cur->_M_next = __next->_M_next;
_M_delete_node(__next);
__next = __cur->_M_next;
--_M_num_elements;
}
}
}
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
::_M_erase_bucket(const size_type __n, _Node* __last)
{
_Node* __cur = _M_buckets[__n];
while (__cur != __last) {
_Node* __next = __cur->_M_next;
_M_delete_node(__cur);
__cur = __next;
_M_buckets[__n] = __cur;
--_M_num_elements;
}
}
//清空hashtable，但是没有释放bucket vector空间
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>::clear()
{
for (size_type __i = 0; __i < _M_buckets.size(); ++__i) {//遍历每个桶子
_Node* __cur = _M_buckets[__i];//当前节点为桶子的第一个节点
while (__cur != 0) {//遍历桶子维护的链表，并释放每个链表节点
_Node* __next = __cur->_M_next;
_M_delete_node(__cur);
__cur = __next;
}
_M_buckets[__i] = 0;//桶子链表为空
}
_M_num_elements = 0;//链表节点数为0
}
//拷贝hashtable对象
//实现机制：首先把原始对象清空，再把空间变成被复制对象__ht一样的大小
//最后把__ht的内容复制到目标对象
template <class _Val, class _Key, class _HF, class _Ex, class _Eq, class _All>
void hashtable<_Val,_Key,_HF,_Ex,_Eq,_All>
::_M_copy_from(const hashtable& __ht)
{//_M_buckets是vector维护的，可以调用vector的成员函数
_M_buckets.clear();//清空线性表调用vector::clear()
_M_buckets.reserve(__ht._M_buckets.size());//把线性表内存空间变成与__ht一样大
//在_M_buckets vector尾端插入size个元素，初始值为空指针
//注意：此时_M_buckets vector为空，即尾端也是起始端
_M_buckets.insert(_M_buckets.end(), __ht._M_buckets.size(), (_Node*) 0);
__STL_TRY {//开始复制操作
for (size_type __i = 0; __i < __ht._M_buckets.size(); ++__i) {
//复制vector的每一个元素(是指向hashtable节点指针)
const _Node* __cur = __ht._M_buckets[__i];
if (__cur) {
_Node* __copy = _M_new_node(__cur->_M_val);
_M_buckets[__i] = __copy;
//针对每一个hashtable节点对应的list，复制list的每一个节点
for (_Node* __next = __cur->_M_next;
__next;
__cur = __next, __next = __cur->_M_next) {
__copy->_M_next = _M_new_node(__next->_M_val);
__copy = __copy->_M_next;
}
}
}
_M_num_elements = __ht._M_num_elements;//更新节点个数
}
__STL_UNWIND(clear());
}
__STL_END_NAMESPACE
#endif /* __SGI_STL_INTERNAL_HASHTABLE_H */
// Local Variables:
// mode:C++
// End:
5.7.7 hash functions
在SGI STL中hash表的实现是采用拉链法，其中用到了哈希函数，哈希函数的作用是计算元素位置的函数，一般哈希值是键值对桶子数取余。在SGI STL提供的哈希函数是有限的，只支持特定的元素类型，若用户需要使用其他类型的哈希函数，则必须自行定义。定义的时候注意一下几点：
使用struct，然后重载operator().
返回是size_t
参数是你要hash的key的类型。
函数是const类型的。
例如定义string类型的哈希函数：
struct str_hash{
size_t operator()(const string& str) const
{
unsigned long __h = 0;
for (size_t i = 0 ; i < str.size() ; i ++)
__h = 5*__h + str[i];
return size_t(__h);
}
};
hash函数源码剖析，源码位于stl_hash_fun.h中
#ifndef __SGI_STL_HASH_FUN_H
#define __SGI_STL_HASH_FUN_H
#include <stddef.h>
__STL_BEGIN_NAMESPACE
//hash function 是计算元素位置的函数
//这些函数可以对hashtable进行取模运算
//这是hashtable所提供的散列函数是取模运算决定的
/*
SGI hashtable以下有限的定义类型：
struct hash<char*>
struct hash<const char*>
struct hash<char>
struct hash<unsigned char>
struct hash<signed char>
struct hash<short>
struct hash<unsigned short>
struct hash<int>
struct hash<unsigned int>
struct hash<long>
struct hash<unsigned long>
不在这里定义的类型，不能使用，若用户想要使用，则必须自己定义。例如：string，double，float
*/
/*定义自己的哈希函数时要注意以下几点：
[1]使用struct，然后重载operator().
[2]返回是size_t
[3]参数是你要hash的key的类型。
[4]函数是const类型的。
*/
template <class _Key> struct hash { };
//对const char* 提供字符串转换函数
inline size_t __stl_hash_string(const char* __s)
{
unsigned long __h = 0;
for ( ; *__s; ++__s)
__h = 5*__h + *__s;
return size_t(__h);
}
__STL_TEMPLATE_NULL struct hash<char*>
{
size_t operator()(const char* __s) const { return __stl_hash_string(__s); }
};
__STL_TEMPLATE_NULL struct hash<const char*>
{
size_t operator()(const char* __s) const { return __stl_hash_string(__s); }
};
//下面的hash函数都是直接返回原值
//对于char,unsigned char,signed char,int,unsigned int,
//short, unsigned short, long,unsigned long都只是返回数值本身
__STL_TEMPLATE_NULL struct hash<char> {
size_t operator()(char __x) const { return __x; }
};
__STL_TEMPLATE_NULL struct hash<unsigned char> {
size_t operator()(unsigned char __x) const { return __x; }
};
__STL_TEMPLATE_NULL struct hash<signed char> {
size_t operator()(unsigned char __x) const { return __x; }
};
__STL_TEMPLATE_NULL struct hash<short> {
size_t operator()(short __x) const { return __x; }
};
__STL_TEMPLATE_NULL struct hash<unsigned short> {
size_t operator()(unsigned short __x) const { return __x; }
};
__STL_TEMPLATE_NULL struct hash<int> {
size_t operator()(int __x) const { return __x; }
};
__STL_TEMPLATE_NULL struct hash<unsigned int> {
size_t operator()(unsigned int __x) const { return __x; }
};
__STL_TEMPLATE_NULL struct hash<long> {
size_t operator()(long __x) const { return __x; }
};
__STL_TEMPLATE_NULL struct hash<unsigned long> {
size_t operator()(unsigned long __x) const { return __x; }
};
__STL_END_NAMESPACE
#endif /* __SGI_STL_HASH_FUN_H */
// Local Variables:
// mode:C++
// End: